INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The rapid development of gas exchange abnormalities with severe arterial hypoxemia is a major threat in lung transplantation (1). Such disturbances mostly coincide with pulmonary edema formation as well as radiographic, hemodynamic, and clinical abnormalities identical to the adult respiratory distress syndrome (ARDS). Corresponding alterations of lung fluid balance, vasomotor regulation, and gas exchange are well known to occur after fibrinolytic or operative removal of thromboembolic material from the lung vasculature (2, 3). Impairment of capillary endothelial and alveolar epithelial barrier function is assumed to represent a major underlying event in such circumstances of lung ischemia-reperfusion, and leukocyte-dependent and independent mechanisms of oxidant attack have particularly been implicated in this pathogenetic sequelae (4).

Inhalation of nitric oxide (NO) can improve gas exchange in experimental models of acute lung injury and under clinical conditions of severe ARDS, with regional vasodilation in well-ventilated (NO-accessible) lung areas, and redistribution of blood flow to these areas being suggested as the predominant underlying mechanism (7). In addition, there is increasing evidence that NO may modulate microvascular barrier properties via direct or indirect mechanisms (8). Decreased endogenous lung NO synthesis was noted in lung ischemia-reperfusion (9). Recent studies addressing the influence of NO inhalation during the period of lung reperfusion on pulmonary vascular permeability yielded conflicting results, as both reduction of the leakage response (10) and lack of efficacy (14) or even enhanced protein leakage and edema formation (15, 16) were noted.

In the current study, ischemia-reperfusion was provoked in buffer-perfused rabbit lungs, and the multiple inert gas elimination technique (MIGET) was employed for detailed analysis of the sequence of gas exchange abnormalities. Both severe ventilation-perfusion (A/) mismatch and progressive shunt flow were noted to occur under these conditions. Inhalation of NO throughout the period of reperfusion significantly reduced both mismatch and shunt flow, thus supporting the view that prophylactic administration of this vasodilatory agent might be beneficial for maintenance of gas exchange in reperfused ischemic lungs.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Materials

Sterile Krebs-Henseleit hydroxyethylamylopectine buffer (KHHB), containing 120 mM NaCl, 4.3 mM KCl, 1.1 mM KH ₂PO₄, 23 mM NaHCO₃, 2.4 mM CaCl₂, 1.3 mM MgPO₄, 2.4 g/L glucose as well as 5% (wt/vol) hydroxyethylamylopectine (molecular weight 200,000) as oncotic agent was obtained from Serag-Wiessner (Naila, Germany). Halothane was supplied by Hoechst AG (Frankfurt am Main, Germany). Gas mixtures of sulfur hexafluoride (SF₆), ethane and cyclopropane (10, 20, and 70%; quantified by gas chromatography), and NO (1,000 parts per million [ppm], 300 ppm, and 10 ppm in N₂) were from Messer Griesheim (Frankfurt-Griesheim, Germany). All other biochemicals (diethyl ether and acetone) were purchased from Merck (Darmstadt, Germany). Columns for gas chromatography, already packed with Hayesep Q mesh 80/100 (flame ionization detector) and a 5-Å molecular sieve, were obtained from Chrompack (Frankfurt am Main, Germany).

Isolated Lung Model

The perfused lung model has been previously described in detail (17). Briefly, rabbits of either sex (weighing 2.5 to 3.0 kg) were anesthetized with intravenous ketamine/xylazine and anticoagulated with heparin (1,000 U/kg). Tracheostomy was performed, and the animals were ventilated with room air, using a Harvard respirator (cat/rabbit ventilator 6025; Hugo Sachs Elektronik, March Hugstetten, Germany). After midsternal thoracotomy, catheters were placed into the pulmonary artery and the left atrium, and perfusion with KHHB was started. Sterilized perfusion circuit tubing was used throughout. For washout of blood, perfusate was initially not recirculated and the lungs were removed from the thorax without interruption of ventilation and perfusion. The lungs were placed in a temperature-equilibrated, humidified chamber at 37.5° C, freely suspended from a force transducer for monitoring of organ weight. In a recirculating system the flow was slowly increased to 135 ml/min (total volume 350 ml). Pulmonary venous pressure was set at 1.5 to 2 mm Hg in all experiments. In parallel with the onset of artificial perfusion, warmed and humidified room air supplemented with 5% CO₂ was used for ventilation (tidal volume 30 ml, frequency 11/min). A positive end-expiratory pressure of 1.0 cm H₂O was used throughout. Pressures in the pulmonary artery, the left atrium, and the trachea were registered by means of small-diameter tubing threaded into the perfusion catheters and the trachea and connected to pressure transducers (zero referenced at the hilum). Perfusate samples were taken by use of these tubes. Gas samples were taken from the outlet of a copper expiration gas mixing box. The whole system was heated to 37.5° C, monitored by temperature sensoring in the pulmonary artery catheter and the expiration gas mixing box.

Lungs included in the study had a homogeneous white appearance with no signs of hemostasis, edema, or atelectasis; pulmonary artery and ventilation pressures in the normal range; and were isogravimetric (lung weight gain < 0.4 g/h) during an initial steady-state period of at least 45 to 60 min. According to these criteria, less than 5% of all lung preparations were discarded before entering the study.

A/ Determination in Isolated Lungs by MIGET

The continuous A/ distributions were determined by MIGET as described by Wagner and coworkers (18), which has been adapted to assess gas exchange of blood-free perfused rabbit lungs (17). Briefly, six inert gases (SF ₆, ethane, cyclopropane, halothane, diethyl ether, and acetone) were dissolved in KHHB and continuously infused at a rate of 0.5 ml/min. After an equilibration period of at least 40 min, 10-ml perfusate samples were simultaneously collected from the pulmonary artery and the venous effluent. A corresponding 30-ml gas sample was drawn from the heated expiration gas mixing chamber. Extraction of the gases dissolved in the buffer fluid was carried out by equilibration (40 min) with nitrogen in a shaking water bath (37.5° C). The gas phases after equilibration of the buffer samples, as well as the exhaled gases were analyzed by gas chromatography. Separation and quantification of ethane, cyclopropane, halothane, diethyl ether, and acetone were performed with a gas chromatograph equipped with a flame ionization detector (model 3300; Varian Associates, Palo Alto, CA), using a commercially available Hayesep Q column. SF₆ was measured by a gas chromatograph fitted with an electron capture detector (model 3300; Varian), using a 5-Å molecular sieve. The ratios of arterial to mixed venous partial pressures (retention) and of expired to mixed venous partial pressures (excretion) were calculated for each gas and were plotted against buffer-gas partition coefficients (retention-solubility and excretion-solubility curves). The duplicate samples of each set of the retentions and the excretions were treated separately, resulting in two A/ distributions by least-squares analysis with enforced smoothing, using a computer program graciously supplied by P. D. Wagner. The position of the distribution was also described by the mean A/ ratio for perfusion ( mean) and ventilation (A mean), and their dispersion by the log standard deviation of both perfusion and ventilation (log SD; log SDA). These parameters of dispersion do not take into account either shunt or dead space. The residual sum of squares (RSS) was the result of testing the compatibility of the inert gas data to the derived A/ distribution by the least-squares method. An indication of acceptable quality of the A/ distributions is a RSS of 5.348 or less in half of the experimental runs (50th percentile) or 10.645 or smaller in 90% of the experimental runs (90th percentile) (19). In the present study, 79.2% of RSS was less than 5.348 and 96.0% was less than 10.645.

Determination of Pulmonary Capillary Pressure

Pulmonary capillary pressure (Ppc) was determined by the double- occlusion technique (20). Interruption of the arterial and venous flow was simultaneously performed by rapid magnetic valves during the apneic period between ventilation breaths. Pulmonary arterial (Ppa) and left atrial pressures were digitized and sampled with an analog-to-digital converter at a frequency of 20 Hz. Ppc was calculated from a 1.5-s interval as the average of the mean pulmonary arterial and venous pressures 3 s after onset of double occlusion. Total pulmonary vascular resistance was partitioned into arterial (precapillary) and venous (postcapillary) resistance: Ra = (Ppa  Ppc)/; Rv = (Ppc  Ppv)/.

Inhalation of NO

NO was delivered into the inspiratory gas flow by use of a gas mixing chamber (Witt, Witten, Germany) while keeping the inspiratory oxygen and CO₂ concentrations as well as the ventilator settings constant. The level of NO in the inspired gas was controlled by a chemiluminescence detector (UPK 3100; UPK Bad Nauheim, Germany).

Experimental Protocols

A total of 37 isolated lung experiments was performed. After termination of the initial steady-state period, duplicate samples for estimation of baseline A/ relationships were drawn. Thereafter, four baseline double-occlusion maneuvers were performed. In the course of the ischemic period, interventions were performed according to one of the following protocols:

120 min of anoxic ischemia. After 45 to 60 min of normoxic ventilation, time was set at zero and anoxic ischemia was initialized. Perfusion was interrupted, and the arterial and venous catheters were both clamped for maintenance of a positive intravascular pressure during ischemia (Table 1). Lungs were continuously ventilated with an anoxic gas mixture containing 95% N₂ and 5% CO₂. After 120 min, ventilation was switched to normoxic gas mixture, and reperfusion was initialized by gradually increasing the flow rate to 135 ml/min within 30 s. Further A/ measurements were undertaken at 135, 150, 180, and 210 min, and double-occlusion maneuvers were undertaken at 121, 123, 125, 130, 135, 150, 180, and 210 min, respectively

180 min of anoxic ischemia. These experiments were performed accordingly, with an ischemic period extended to 180 min. A/ measurements and double-occlusion maneuvers were undertaken at 195, 210, 240, and 270 min, and at 181, 183, 185, 190, 195, 210, 240, and 270 min, respectively.

Controls. Continuous perfusion and normoxic ventilation were performed for intervals corresponding to the ischemia experiments (210 and 270 min, respectively). No interventions were undertaken. A/ measurements and double-occlusion maneuvers were carried out following the same protocols as used for the ischemic lungs.

Application of NO. After 120 min of anoxic ischemia, 1 ppm, 10 ppm, 50 ppm, or 250 ppm of NO were admixed to the inspiration gas 60 s before the onset of reperfusion. In additional experiments with 180 min of anoxic ischemia, 50 ppm NO was administered 60 s before the onset of reperfusion. In all groups, NO supply was continued until the termination of the experiments. A/ distributions and capillary pressures were determined at time points as detailed previously.

All experiments were terminated after 90 min of reperfusion, or when lung weight gain exceeded 25 g during reperfusion paralleled by alveolar edema formation with foam ascending into the ventilation circuit.

Statistical Analysis

Data are given as means ± SEM. Statistical differences within each group were determined by one-way repeated-measures analysis of variance (ANOVA) followed by post hoc Student-Newman-Keuls test. Differences between groups were tested by one-way ANOVA followed by post hoc Student-Newmans-Keuls test. A p value of < 0.05 was considered significant. All statistical tests were performed using the SigmaStat software package (Jandel Scientific, San Rafael, CA).

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Baseline Conditions

After termination of the steady-state period, all lungs displayed mean pulmonary artery pressure (Ppa) values of 4 to 8 mm Hg and were isogravimetric (lung weight gain < 0.4 g/h). Baseline A/ measurements revealed physiological ventilation-perfusion distributions in all experiments (Tables 2 and 3). Predominant unimodal distribution of perfusion to midrange A/ areas (0.1 < A/ < 10) was observed throughout. Shunt flow (A/ < 0.005) and perfusion to poorly ventilated regions (0.005 < A/ < 0.1) ranged below 3%, and no perfusion of high A/ regions (10 < A/ < 100) was observed. Narrow dispersion of perfusion in the midrange A/ regions was noted (log SD 0.35 to 0.60). Except for the dead space ventilation, augmented due to the tubing connecting the trachea and the ventilator, gas flow was exclusively distributed to midrange A/ areas (0.1 < A/ < 10). The low log SD of ventilation indicated narrow dispersion of ventilation distribution.

Hemodynamics and Edema Formation in Response to 120 and 180 min of Ischemia

Reperfusion after ischemia provoked an immediate increase in Ppa, with peak values ranging between 11 and 15 mm Hg in response to 120 min ischemia (Figure 1), and 11 and 19 mm Hg in response to 180 min of ischemia. This pressor response was partially reversible, with still moderately elevated Ppa levels 15 and 30 min after onset of reperfusion. Analysis of the distribution of pre- and postcapillary vascular resistance showed that the increase in Ppa was mainly caused by enhanced precapillary resistance (Ra), with only moderate elevation of venous vascular resistance (Rv) and Ppc (Figure 2, Tables 4 and 5). In addition to the pressor response, both 120 and 180 min of ischemia provoked progressive edema formation as reflected by the continuous increases of organ weight (Figure 3, Table 5).

View larger version (25K): [in this window] [in a new window]   Figure 1.   Impact of different concentrations of inhaled NO on the Ppa response to reperfusion. At time zero, lungs were rendered ischemic, and reperfusion was started at 120 min. Means ± SEM of n = 4 independent experiments each (n = 5 for 50 ppm NO, respectively); error bars are missing when falling into symbol. # p < 0.05 versus control; & p < 0.05 versus control as well as 1, 10, 50, and 250 ppm NO; § p < 0.05 versus preischemic value.

View larger version (21K): [in this window] [in a new window]   Figure 2.   Impact of different concentrations of inhaled NO on the Ppc response to reperfusion. At time zero, lungs were rendered ischemic, and reperfusion was started at 120 min. Means ± SEM of n = 4 independent experiments each (n = 5 for 50 ppm NO, respectively); error bars are missing when falling into symbol.

View larger version (21K): [in this window] [in a new window]   Figure 3.   Impact of different concentrations of inhaled NO on lung weight gain in reperfused rabbit lungs. At time zero, lungs were rendered ischemic, and reperfusion was started at 120 min. Means ± SEM of n = 4 independent experiments each (n = 5 for 50 ppm NO, respectively); error bars are missing when falling into symbol. The ischemia groups were all significantly different from nonischemic controls (p < 0.05). The 1, 10, 50, and 250 ppm NO groups were all different from the 0 ppm NO group (p < 0.05). $p < 0.05 versus 1 and 250 ppm NO.

Influence of Inhaled NO on Pulmonary Hemodynamics and Edema Formation after 120 min of Ischemia

In all concentrations used, inhaled NO significantly attenuated the Ppa and Ra rise in response to reperfusion (Figure 1, Table 4). This effect was least prominent for 1 ppm NO, and maximal efficacy was noted for 10 ppm NO; increasing the NO concentration to 50 and 250 ppm did not result in a further suppression of the reperfusion-related Ppa and Ra increase. There was only a very moderate increase in Rv data upon reperfusion in the presence of NO, which was comparable to the absence of NO. Similarly, only minor, nonrelevant changes in Ppc data were observed (Figure 2, Table 4).

Inhalation of exogenous NO significantly attenuated edema formation during reperfusion in a dose-dependent manner (Figure 3), thereby enabling prolonged observation periods after onset of reperfusion. Maximal and equal efficacy was noted for 10 and 50 ppm NO, with 1 and 250 ppm NO being significantly less potent.

Influence of Inhaled NO on Pulmonary Hemodynamics and Edema Formation after 180 min of Ischemia

The inhalation of 50 ppm NO after 180 min of anoxic ischemia neither reduced the pressor response nor the rapid lung weight gain upon reperfusion (Table 5). All variables measured were comparable to lungs reperfused after 180 min of ischemia in the absence of NO.

A/ Distribution after 120 and 180 min of Ischemia

120 min of anoxic ischemia caused a severe deterioration of A/ matching upon reperfusion (Figures 4-6, Table 2). Shunt flow increased progressively to a mean of > 15% within 15 min, accompanied by a significant perfusion of low A/ areas as well as some enhanced flow to high A/ regions. The mean shifted to the left, and the dispersion of perfusion (log SD) increased more than 3-fold. In the further course, perfusion of low A/ areas declined, paralleled by a further increase of shunt flow to a mean of > 30% thirty min after onset of reperfusion. Dead space and high A/ ventilation were slightly elevated (data not given in detail). The dispersion of ventilation was broadened, and the mean A increased. Comparable changes of A/ matching were noted after 180 min of ischemia. The broadening of ventilation distribution was more prominent under these conditions as compared with 120 min of ischemia.

View larger version (28K): [in this window] [in a new window]   Figure 4.   Examples of A/ distribution in isolated lungs under different experimental conditions. ( Top left) Control lungs after 135 min of perfusion without ischemia. (Top right) Lungs after 120 min of anoxic ischemia and 15 min of reperfusion. (Bottom left) Lungs after 120 min of anoxic ischemia and 15 min of reperfusion in the presence of 10 ppm NO. (Bottom right) Lungs after 180 min of anoxic ischemia and 15 min of reperfusion.

View larger version (22K): [in this window] [in a new window]   Figure 5.   Impact of different concentrations of inhaled NO on shunt flow in reperfused rabbit lungs. At time zero, lungs were rendered ischemic, and reperfusion was started at 120 min. Shunt is expressed in percent of total perfusion. All data were obtained by MIGET. Means ± SEM of n = 4 independent experiments each (n = 5 for 50 ppm NO, respectively); error bars are missing when falling into symbol. #p < 0.05 versus controls; & p < 0.05 versus controls as well as 10, 50, and 250 ppm NO, § p < 0.05 versus preischemic value.

View larger version (22K): [in this window] [in a new window]   Figure 6.   Impact of different concentrations of inhaled NO on the appearance of low A/ areas (areas with a ventilation-perfusion ratio between 0.005 and 0.1), given in percent of total perfusion. At time zero, lungs were rendered ischemic, and reperfusion was started at 120 min. All data were obtained by MIGET. Means ± SEM of n = 4 independent experiments each (n = 5 for 50 ppm NO, respectively); error bars are missing when falling into symbol. # p < 0.05 versus controls; & p < 0.05 versus controls as well as 10, 50, and 250 ppm NO, § p < 0.05 versus preischemic value.

Influence of NO on the A/ Distribution after 120 min of Ischemia

Inhalation of NO caused a dose-dependent attenuation of the reperfusion-related deterioration of A/ matching. The rank order of efficacy was 10 ppm NO  50 ppm NO > 250 ppm NO > 1 ppm NO. In the presence of 10 and 50 ppm NO, shunt flow and perfusion of low A/ areas never surpassed 8% and 5%, respectively, until the end of experiments 90 min after onset of reperfusion (Figures 5 and 6). The leftward shift of mean perfusion was reduced, with more narrow distribution of perfusion as compared with the absence of NO (Table 2). Similarly, the rightward shift as well as the broadening of ventilation dispersion were significantly reduced in the presence of NO.

Influence of NO on the A/ Distribution after 180 min of Ischemia

Inhaled NO (50 ppm) did not improve the deterioration of the A/ matching after 3 h of anoxic ischemia. All variables assessed by the MIGET technique did not differ from the corresponding data in the untreated ischemic group (Table 3).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Reperfusion of rabbit lungs after 120 and 180 min of warm ischemia provoked a partially reversible initial vasopressor response, mostly due to increased precapillary vascular resistance, and progressive pulmonary edema formation. Most impressively, severe abnormalities of gas exchange were observed immediately after onset of reperfusion. These included major ventilation-perfusion mismatch with appearance of low A/ areas and broadening of both ventilation and perfusion dispersion in the midrange A/ regions. In addition, progressive and marked shunt flow was noted. Inhalation of NO throughout the reperfusion period significantly reduced the initial transient pressor response and the lung edema formation, and markedly attenuated the severe gas exchange abnormalities in lungs undergoing 120 min of warm ischemia. This protective effect was more prominent for 10 and 50 ppm NO than for 1 and 250 ppm NO. In contrast, the overwhelming abnormalities in gas exchange and fluid balance provoked by a 3-h period of warm ischemia were not prevented by NO inhalation.

As anticipated, a partially reversible increase in Ppa occurred upon onset of reperfusion after 120 and 180 min of warm ischemia. The monitoring of the microvascular pressure during this pressor response clearly suggested that it was mostly the result of an increase in precapillary vascular resistance. It is unlikely that a persisting acute hypoxic pulmonary vasoconstriction contributed to the rise in Ppa, as a normal alveolar P O₂ was reestablished 1 min before onset of reperfusion. Thromboxane has been noted in previous studies to be a major contributor to this immediate hypertensive response in rabbit lungs undergoing ischemia-reperfusion (21). It is well compatible with this suggestion that the vasorelaxant agent NO, in concentrations between 10 and 250 ppm, blocked the pressor response during reperfusion by more than 50%. The time course of the very moderate microvascular pressure elevation was, however, not relevantly affected by NO.

In addition to the transient pressor response, progressive edema formation represented a consistent feature in the reperfusion period. As the microvascular pressure values were only very moderately and transiently elevated, this leakage response may not be caused by enhanced hydrostatic capillary fluid filtration, but must be ascribed to an increase in lung capillary permeability. Preceding investigations in the current model have indeed demonstrated markedly augmented capillary filtration coefficients in the reperfusion period (21) in line with previous observations of impaired endothelial barrier function in lung ischemia-reperfusion (10, 12, 15, 22). Such permeability increase occurred though maneuvers of biophysical "protection" of the lung parenchyma were performed. First, anoxic ventilation was performed, as ventilation-dependent dynamic mechanical forces were previously described to attenuate ischemia-related lung injury (21, 23). Second, a positive intravascular pressure was maintained throughout the ischemic period, which was recently noted to reduce the ischemia- reperfusion-related leakage response (21, 24). The underlying mechanisms of impaired endothelial barrier function in lung ischemia-reperfusion are still not fully settled. Different types of oxidant injury, involving oxygen radical formation in recruited neutrophils, macrophages, and the endothelial cells themselves (6, 25, 26) as well as disturbances of the endothelial calcium homeostasis with subsequent actin-dependent endothelial contraction and interendothelial gap formation (27) have been implicated in this phenomenon.

Several of these mechanisms may be affected by NO, and inhalation of this agent indeed suppressed the leakage response to ischemia-reperfusion in the current study. In endothelial monolayers in vitro, NO was demonstrated to inhibit actin-dependent intercellular gap formation by elevation of intracellular cyclic guanosine monophosphate (cGMP) levels and activation of cGMP-dependent kinases (28). Moreover, NO may inhibit the release of reactive oxygen species from neutrophils (29) known to reside in large quantities in the pulmonary microvasculature (30) and to contribute to ischemia- reperfusion-related lung injury even under conditions of buffer perfusion (5). In this context, NO-induced reduction of leukocyte-endothelial adhesion may be additionally operative (31). The presently observed reduction of lung edema formation, being maximal at 10 to 50 ppm NO, is well in line with reports on a protective effect of NO inhalation on reperfusion-related leakage in isolated neonatal piglet (10) and rat lungs (11). Moreover, protection by NO was observed in reperfused isolated rat lungs treated with NO donors (32). In contrast, NO inhalation failed to provide protection or even enhanced the leakage response in different rat lung transplantation models (14). Interestingly, in one of these models a cGMP analog reduced the reperfusion injury whereas inhaled NO was ineffective (9, 14), thus pointing at possible disadvantageous effects of NO besides its stimulation of the soluble guanylate cyclase. Differences in the susceptibility to NO toxicity might be one explanation for the above-described inconsistent response of reperfusion-related lung vascular leakage to NO inhalation.

The current study is the first to assess the gas exchange abnormalities in response to pulmonary ischemia-reperfusion by MIGET analysis in detail. In summary, the changes were characterized by a redistribution of perfusate flow from normal A/ areas to low A/ regions and, progressively, to shunt areas, accompanied by a mismatch of ventilation-perfusion distribution also in the midrange A/ regions. Two mechanisms appear to offer the most favorable explanation for these severe gas exchange abnormalities. First, the progressive leakage of fluid may result in first partial and subsequently complete edema filling of alveolar spaces, and the initial predominance of low A/ areas with subsequent decline of these regions and progressive rapidity in the appearance of shunt areas is well in accord with such a notion. It may be assumed that it is not only the alveolar fluid filling per se, but also subsequent alterations in surfactant, giving way to alveolar instability that will result in progressive alveolar collapse in the most affected areas. Second, the generation of vasoconstrictor agents by the reperfusion-related inflammatory events may result in severe A/ mismatch including the appearance of major shunt flow, even in the absence of significant lung edema formation. This was previously demonstrated in perfused rabbit lungs undergoing challenge with bacterial toxins such as staphylococcal -toxin and Escherichia coli hemolysin (33, 34) and is reproduced by direct infusion of the stable thromboxane analogue U-46619 in this model (35). It is suggested that the provocation of vasoconstrictor formation in areas that are physiologically well perfused to match the alveolar ventilation distribution results in redistribution of flow to badly or even nonventilated regions, putatively distributed in a patchy pattern, which are spared from perfusion by hypoxic vasoconstriction at normal pulmonary artery pressures. The vasomotor activities of locally arising inflammatory agents thus interfere with or even predominate over the physiologic pattern of perfusion distribution, which aims to match the ventilation distribution by hypoxic vasoconstriction and normoxic vasodilation. Local generation of vasoactive mediators, not regulated by the pattern of alveolar oxygen content, may also well explain the mismatch of ventilation and perfusion in the midrange A/ areas (increase in log SD and log SD A; countercurrent shift of mean and mean A). Overall, the combination of both pathogenic events, vascular leakage and inflammatory vasoconstrictor generation, may well explain the profile of severe gas exchange abnormalities encountered in response to both 120 and 180 min of warm ischemia.

Inhalation of NO throughout the reperfusion period caused a marked improvement of gas exchange, with 10 and 50 ppm NO being significantly more potent than 1 and 250 ppm NO. The efficacy of the inhaled vasodilator turned out to hold true for all aspects of the A/ mismatch described previously, i.e., for the transient appearance of low A/ areas, the progressive formation of shunt flow and the mismatch encountered within the midrange A/ areas. Three aspects may underlie this protective effect of NO in terms of gas exchange. First, edema formation was attenuated in the presence of NO, thereby reducing the impact of this event on the appearance of A/ mismatch and shunt flow. Second, the Ppa rise was reduced, thus mitigating the Ppa-driven redistribution of flow to badly or nonperfused lung regions. Third, the preferential or even exclusive distribution of NO to well ventilated lung regions has been documented to result in regional vasodilation in these areas in various models of acute lung injury (7), thereby supporting the matching of perfusate flow to the distribution of ventilation. A combination of these mechanisms may well explain the impressive amelioration of gas exchange noted in the presence of NO in the lungs under going 120 min of ischemia. In contrast, after 180 min of ischemia all pathogenic events were evidently so overwhelming that even the optimal dosage of NO failed to have any beneficial effect.

Interestingly, the efficacy of NO in reducing edema formation and improving gas exchange was noted to possess bell-shaped dose dependency, with the high dose of 250 ppm being significantly less effective than 10 and 50 ppm. Such a finding may partly be explained by some loss of selectivity of regional vasodilation, as high NO concentrations might "spill over" to badly or nonventilated areas, thus causing vasorelaxation also in these regions. Most probably, however, nonguanylate-cyclase-related "toxic" effects of NO must be assumed to underlie this observation. High concentrations of NO may well out-compete superoxide dismutase for superoxide, resulting in the generation of the highly reactive peroxynitrite, and the formation of NO₂ from NO and oxygen is critically dependent on the NO concentration (7). Both reaction products have been implicated in severe cellular damage, with mechanisms including membrane lipid peroxidation and DNA strand breaks. NO toxicity might be favored in the current model by the absence of hemoglobin, known to rapidly bind NO once entering the intravascular space.

In conclusion, severe gas exchange abnormalities including both major A/ mismatch and shunt flow were found to appear rapidly in reperfused ischemic rabbit lungs, along with a transient vasopressor response and progressive pulmonary edema formation. Medium doses of inhalative NO, administered from onset of reperfusion, significantly protected from vascular leakage, A/ mismatch, and shunt flow, whereas high NO doses were less effective, presumably because of toxic side effects. Prophylactic administration of low concentrations of NO might thus be considered for maintenance of gas exchange in lung transplantation.